An efficient privacy-preserving data publishing in health care records with multiple sensitive attributes

Emerging advanced technologies in the e-health sector have improved the industry, helping to invent new medicine, reduced cost, better-quality consultation, and treatment. The improvements have originated from the patient record analysis that is shared between the health care provider and researchers. Due to this collaboration, there is an emerging security and privacy requirement in the health care sector. Privacy-preserved data publishing plays a significant role in preventing the disclosure of individual identity. Many privacy-preserving models have been proposed earlier for data publishing. However, they still suffer from few drawbacks i) cannot be applied on 1: M dataset ii) no trade-off between privacy and utility. To satisfy the multi-record privacy and trade-off between classification utility and privacy, we proposed Quasi-identifier Bucket-Individual Multi-Sensitive Attribute Bucket (QIAB-IMSB) algorithm. The main focus of our work is to anonymize the multi-valued record of an individual. The initial phase of our work is to merge the multi-record of an individual. Next, we implement vertical partitioning and apply k-anonymity for quasi-identifier bucket (QIAB) and (k, l) - diversity for individual multi-sensitive attribute bucket (IMSB). Our proposed approach mainly focuses on preventing the sensitive linking attack by implementing hierarchical generalization taxonomy. Our experimental results proved that there is an improved privacy and classification utility in privacy-preserving data publishing of health care records.